ABSTRACT
Thought-controlled electrical wheelchairs have revolutionized the care of paraplegic people. However, these wheelchairs cost several thousands of dollars, and thus, may not be widely available to all potential users. This pilot project aimed at identifying an approach to convert any electric wheelchair into an inexpensive (∼$600) thought-controlled wheelchair.
This project used the brain–computer interface (BCI) technology to interpret the brain waves of a patient using electroencephalography (EEG) and to convert them to electrical signals to operate devices, such as wheelchairs. This innovation can be attached to any electric wheelchair with a joystick without damaging or permanently modifying the wheelchair.
This pilot project established communication between brain waves and a wheelchair using inexpensive BCI technology to convert an electric wheelchair into an EEG-controlled wheelchair.
BCI is an innovative technology that can be implemented in applications across healthcare, and thus, can impact the lives of patients. Our study demonstrated that an inexpensive wheelchair attachment could be created to convert a conventional wheelchair to a BCI wheelchair. This innovation in principle would allow many patients to be able to take advantage of the BCI technology by making it affordable for everyone.
INTRODUCTION
According to the World Health Organization, 65 million people globally use a wheelchair daily, 8% of whom are paralyzed.[1] Paraplegic wheelchair users cannot control the joysticks of standard electrical wheelchairs, and thus, require assistance to move around. Because of this dependency, disabled people are five times more susceptible to mental distress than people without disabilities.[2–4]
A brain–computer interface (BCI) can be incorporated in many potential applications for disabled persons (e.g., robotic limbs, smart houses, and transportation systems).[5] Its principle is based on reading brain waves and translating these signals into actions and commands that can control a computer. A BCI can make paralyzed people more independent by increasing the accessibility of technologies, by providing an alternative communication channel between the human brain and a computer. This channel does not depend on the normal output channels of the peripheral nerves and muscles of the brain.[6,7]
BCI-controlled wheelchairs are electrical for ease of use and are available in the market; however, they are very expensive (i.e., costing thousands of dollars). The innovation described in this paper is a different approach that has the potential to lower the current price by 90%, and therefore, can contribute to increased usage of thought-controlled wheelchairs in the future.[8,9] The impact of this technology could be extensive, as it could significantly improve the quality of life of approximately 10 million paralyzed people by enabling those who cannot speak or move their limbs to move around without assistance.
A BCI uses brain signals to collect information regarding the intentions of users. The technique relies on measuring brain signals and converting them into electrical signals. Significant research has been conducted about the use of BCI technology to assist physically handicapped patients. However, none of these solutions is completely user-friendly and has some limitations such as high cost, usability, range, and response time. Some of the prior solution attempts are listed in Supplemental Appendix A (available online). Most current BCIs obtain the relevant information from brain activity using electroencephalography (EEG).[10]
EEG is a noninvasive technology that measures brain activity by recording the electrical impulses generated in various regions of the brain using electrodes placed over the scalp. EEG was first used in 1929 by Hans Berger, who recorded brain activity beneath a closed skull and reported changes during different states.
EEG-based BCI has limitations regarding speed, accuracy, and ease of use due to which they have not been easy to use. Currently, BCI is primarily being used in research, medicinal, and clinical contexts. BCI can be implemented using different approaches known as “paradigms.” A paradigm can be understood to be made up of two parts:
What the user is supposed to do mentally
What is presented to the user
There are some commonly used paradigms, such as P300, Steady State Visually Evoked Potentials (SSVEP), and Motor Imagery, as explained in Table 1.[11]
The goal of this project was to build a chair that can take the users where they want to go without physical exertion or assisted help. The chair should be able to take the users to a specific location without the use of hands or any other body part. We used the P300 paradigm to generate EEG signals from different brain areas. To use the P300 paradigm, electrode sensors are used to acquire EEG signals from different brain areas. These electrodes are placed over different regions of the brain to measure the brain waves generated by specific tasks, such as jaw clenching, blinking eyes, and focusing on a specific picture or characters. Computer software is used to amplify these signals and convert them into electrical outputs, which may be used to control a mechanical device, such as a wheelchair robot.
METHODS
The research was conducted as part of an educational project aimed at understanding basic BCI principles and demonstrating technical concepts. All experiments were conducted at the innovation laboratories at DiscoverSTEM from January 2021 to September 2022. Ethical approval was not required, as the BCI is a noninvasive, observational technology that records brain signals through commercially available EEG headsets. The study analyzed anonymous data to minimize privacy concerns and comply with data protection regulations. The study did not involve testing any medical or therapeutic interventions. All study participants were fully informed about the research objectives and procedures, they voluntarily agreed to participate, and parental consent was obtained for all minors involved.
Functional Requirements
The wheelchair needs to move and stop according to a person’s thought to allow travel for those who are incapable. It would need to withstand basic household terrain (carpet, tile, or wood floors) and the weight of the person in the wheelchair.
The targeted wheelchair users need a chair that can take them where they want to go without physical exertion or assisted help and without the use of a joystick. The user needs to only use their brain to control the wheelchair and move it accordingly. An electric wheelchair was needed that could be molded with a raspberry pi, an algorithm that can work with the raspberry pi and also the BCI, a brain cap that that could have electrodes attached to it and could give the alpha waves, which then could be read by algorithm and then be manipulated to go left, right, forward, or backward, and a design that is connected by raspberry pi and can move the joystick depending on the signals coming from the raspberry pi.
Key elements:
An electric wheelchair that can be molded with a raspberry pi
An algorithm that can work with the raspberry pi and also the BCI
A brain cap that that can have electrodes attached to it and can give the alpha waves that then can be read by algorithm and then be manipulated to go left, right, forward, or backward
A design that is connected by a raspberry pi and can move the joystick depending on the signals coming from the raspberry pi
System Breakdown of the Prototype
The prototype consists of four main systems:
System A: An EEG headset to measure brain signals and convert them to electrical signals using electrodes, amplifiers, A/D converter, and a recording device.
System B: Computer to process raw EEG signals from the EEG system and pass them to the raspberry pi via a UDP protocol.
System C: The initial prototype was created using a Sun founder PiCar-S1 in order to incorporate these three subsystems:
Line Following System
Automatic Anti-Collision System
P300 Brain Controller System
System D: Servo Wheelchair Attachment to serve as the “hand” that moves the joystick on the wheelchair. The attachment consists of a servo motor and a 3-dimensional printed fork-shaped configuration that will move with the motor and push the joystick on the wheelchair either forward or backward without the use of any external forces.
A detailed presentation of each of the systems is provided in Supplemental Appendix B (available online).
Experimental Procedure
The Unicorn Hybrid Black application suite was installed. The licenses of the Unicorn Recorder, Unicorn Speller were activated. Bluetooth was activated, and the headset was connected to the computer (Apple, MacBook Air 2020). There were 10 test subjects who participated in the experiment. Each participant wore the Unicorn Hybrid Black EEG headset and started the Unicorn Recorder. A gel was applied to ensure good contact of the electrodes with the scalp and appropriate signal quality. A raspberry pi was connected to the Unicorn Speller software using a screen and internet connection. The raspberry pi was attached to the side of the wheelchair. The wire from a servo motor was connected to the pins on a circuit board.
Calibration
The brain signals were calibrated by showing a series of pictures and characters on the Unicorn Speller Board to each test participant. The participants concentrated on specific characters to select them mentally. The user selected at least four pictures. The system randomly displayed a series of letters, and the user focused on selected letters. After the selected pictures flashed 50 times, the calibration was completed. Finally, copy spelling (matching letters with the selected pictures) was performed to ensure that the calibration was accurate.
Testing
With the wheelchair and Unicorn Speller connected, the participants selected a target command (“L” for left or “R” for right) by concentrating on the corresponding letter and counting the number of flashes.
The system was coded such that based on the participant selection, the computer would communicate with the raspberry pi to send the corresponding output. The raspberry pi would then move the servo motor to turn the wheelchair accordingly.
If the participant selected “L,” then the expected outcome would be for the wheelchair to turn to the left.
If the participant selected “R” then the expected outcome would be for the wheelchair to turn to the right.
Each participant completed 100 trials, 50 each for the L and R commands. The test parameters are given in Table 2.
RESULTS
The experimental test results are presented in Table 3. The average success rate across all participants was 82.8% (828 successful movements out of 1000 trials). The success rate varied among participants, from 79–87%. These results indicate that the BCI system was generally effective in translating thought commands into wheelchair movements.
DISCUSSION
The fairly high success rates in this study demonstrates the potential of BCI technology in assistive mobility applications. Brain signals reflect the activity occurring in different regions of the brain. The brain generates brain waves of different frequencies based on the task being performed. When patients are disabled because of neuromuscular disorders, their brains are still active, but the brain signals are not transmitted to the body parts owing to injury to the spine or neurons. A BCI provides a channel between the brain and body parts by bypassing the spinal cord transmission mechanism.[12]
There have been several studies and research conducted on BCI technology and its applications to improve patient outcomes. Several innovations in BCI technology aim to enhance mobility for individuals with disabilities through thought-controlled wheelchairs. A 2013 patent allows users to control wheelchair direction and speed using brain signals, although it faces challenges like limited range and usability for severely disabled patients. Another design, targeting patients with amyotrophic lateral sclerosis, uses P300 detection and path guidance, but is limited to indoor use because of low response times. Other systems include a complex path-planning algorithm and alpha wave detection, with strengths in ease of use and monitoring capabilities but weaknesses in navigation and slow response times. Some of these existing solutions are discussed in detail in Supplemental Appendix A. However, most resulting products from this research are expensive and inaccessible to patients owing to their high cost.[13]
To address this problem, we developed an inexpensive attachment that can easily convert any power wheelchair to an EEG-controlled wheelchair (cost ∼$600 US Dollars). This innovation can be easily attached to any electric wheelchair with a joystick without damaging or permanently modifying the wheelchair. The system is also portable because the headset and laptop are connected via Bluetooth and the laptop and raspberry pi are connected via the user datagram protocol. The implications of this novel approach for controlling wheelchair movement are also important, as this attachment—once refined—could improve the mobility and quality of life of people who cannot speak or move their limbs. Future research and development could focus on enhancing the system's precision to increase reliability of the BCI technology for wheelchair users. Further improvements of the current system include potential integration of manual control of the wheelchair and omnidirectional movements with options such as forward, backward, left, and right movements.
The BCI technology can add further features to make wheelchairs more user-friendly, such as incremental path-planning algorithms to navigate a map built using light detection and ranging (LiDAR), although this will incur additional costs. A wheelchair can include a rotating camera and LiDAR sensor to identify objects in the path and can avoid collisions by building an automatic brake mechanism. Moreover, a wheelchair can follow a predetermined path and navigate to the destination of a user faster using a light-following sensor.[14]
BCI applications are not restricted to medical science and healthcare but can be incorporated into gaming and entertainment, neuro-ergonomics, and smart environments.[15–17] For example, BCI assistive robots are expected to offer support to disabled users in daily and professional life by building hands-free applications, such as mind-controlled smart lighting and heating systems. They can also help in monitoring the mental and emotional states of an individual by tracking their brain activity.[6,15–17] All of these potential extensions of this proof-of-principle study need to be tested thoroughly, as each additional step also has the potential to increase complexity and associated costs.
CONCLUSION
This pilot project established communication between brain waves and a wheelchair using inexpensive BCI technology, which can convert an electric wheelchair into an thought-controlled wheelchair.
Supplemental Material
Supplemental material is available online with the article.
References
Competing Interests
Source of Support: None. Conflict of Interest: None.